Understanding network behavior patterns of bus Wi-Fi users using surfing data

Nowadays, more and more companies intend to combine public transport system with the mature Wi-Fi technology to bring more passengers value-added services during their travel. It is critical to mobile service market and city management by identify the user's behavior patterns. In this paper, the author firstly analyzes the time regularity of bus Wi-Fi network behavior. Then from the perspective of users and websites, analyzes the behavior features from the number of records, flow, website numbers and the number of users. Based on the results of those feature analysis, we get the most important 1000 websites and use the app classification method, extract 16 features, using TF-IDF to improve the clustering algorithm, successfully divide network behavior patterns into four categories. Finally, the author compares and analyzes the network behavior patterns in different scenarios.